Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A method performed by a computing system for detecting pathological anomalies in a digital pathology image, comprising: providing a digital pathology image to the computing system; providing an identification model using a machine learning algorithm trained on a plurality of annotated digital pathology images; and analyzing the digital pathology image using an identification module arranged on the computing system, wherein the identification module uses a machine learning module to execute: recognizing an object containing an abnormal image pattern using the identification model loaded in said identification module; and identifying whether the abnormal image pattern corresponds to a pathological anomaly using the identification model, wherein providing the identification model comprises: generating at least one first digital image of a tissue slide comprising stained morphological features of the pathological anomaly, wherein the tissue slide is stained with: a first histochemical dye, or at least one biomarker and at least one fluorescence dye, or at least one fluorescent biomarker; wherein said tissue slide is stained with multiple fluorescent dyes which emit light at different wavelengths; generating at least one second digital image of the same tissue slide, wherein the tissue slide is stained with a second histochemical dye, said second histochemical dye is H&E; annotating the at least one second digital image of the tissue slide by layering the at least one first digital image and the at least one second digital image over another; and providing the layered, annotated images to the machine learning module and training the identification model using different morphological features of the pathological anomalies to obtain a classifier, wherein the step of analyzing the digital pathology image comprises analyzing the multiple fluorescent dyes on the same tissue slide using the classifier.
This invention relates to a method for detecting pathological anomalies in digital pathology images using machine learning. The method addresses the challenge of accurately identifying abnormal tissue patterns in pathology images, which is critical for disease diagnosis and treatment planning. The system processes a digital pathology image by first providing an identification model trained on annotated digital pathology images. The model is trained using a combination of histochemical and fluorescent staining techniques. Specifically, the method involves generating at least one digital image of a tissue slide stained with either a histochemical dye, biomarkers with fluorescence dyes, or fluorescent biomarkers. The tissue slide is stained with multiple fluorescent dyes that emit light at different wavelengths. Additionally, a second digital image of the same tissue slide is generated using a second histochemical dye, such as hematoxylin and eosin (H&E). The first and second images are layered and annotated to create a training dataset. The machine learning model is trained on these annotated images to recognize morphological features of pathological anomalies. During analysis, the system uses the trained model to detect objects containing abnormal patterns and determine whether these patterns correspond to pathological anomalies. The method leverages the multi-wavelength fluorescence staining to enhance the accuracy of anomaly detection. This approach improves the reliability of digital pathology analysis by combining multiple staining techniques with machine learning.
2. The method according to claim 1 , wherein identifying whether the abnormal image pattern corresponds to a pathological anomaly comprises: classifying the abnormal image pattern using a classifier in the identification model to classify the abnormal image pattern in accordance with at least two classes; and determining whether the abnormal image pattern corresponds to a pathological anomaly based on the classification.
This invention relates to medical image analysis, specifically detecting and classifying abnormal image patterns to identify pathological anomalies. The method addresses the challenge of accurately distinguishing between benign and pathological abnormalities in medical imaging, such as X-rays, MRIs, or CT scans, where false positives or negatives can lead to misdiagnosis. The process involves analyzing medical images to detect abnormal patterns, then classifying these patterns using a trained machine learning model. The classification step involves a classifier within the identification model that categorizes the abnormal pattern into at least two predefined classes. Based on this classification, the system determines whether the pattern corresponds to a pathological anomaly. The classifier is trained to differentiate between normal variations, benign abnormalities, and true pathological conditions, improving diagnostic accuracy. The method ensures that detected abnormalities are not only identified but also evaluated for clinical significance, reducing unnecessary follow-ups for non-pathological findings. The classifier may use features extracted from the image, such as texture, shape, or intensity, to make its determination. This approach enhances the reliability of automated medical image analysis, supporting radiologists in faster and more precise diagnoses.
3. The method according to claim 2 , wherein the at least one class is a cancer type and/or a cancer grade and/or a cancer stage.
This invention relates to medical diagnostics, specifically to methods for classifying biological samples, such as tissue or fluid samples, to determine cancer characteristics. The method involves analyzing a sample to identify at least one class, where the class can be a cancer type, cancer grade, or cancer stage. The classification is based on molecular or cellular features extracted from the sample, which are then compared to reference data to determine the most likely class. This approach helps clinicians assess the severity and progression of cancer, aiding in treatment planning. The method may use machine learning or statistical models to improve accuracy. By categorizing cancer into specific types, grades, or stages, the technique provides a structured way to evaluate tumors, enabling more precise and personalized medical interventions. The invention addresses the need for reliable, automated cancer classification to support diagnostic decisions and improve patient outcomes.
4. The method according to claim 1 , wherein providing the identification model comprises: providing the annotated images to the machine learning module; and training the identification model using different morphological features of the pathological anomalies to obtain the classifier.
This invention relates to a method for training a machine learning-based identification model to detect pathological anomalies in medical images. The method addresses the challenge of accurately identifying and classifying anomalies in medical imaging data, which is critical for early diagnosis and treatment planning. The invention focuses on improving the robustness and accuracy of anomaly detection by leveraging diverse morphological features of pathological anomalies during model training. The method involves providing annotated medical images to a machine learning module, where the annotations highlight regions of interest containing pathological anomalies. The training process utilizes various morphological features of these anomalies, such as shape, size, texture, and spatial distribution, to enhance the model's ability to distinguish between normal and abnormal tissue. By incorporating these diverse features, the trained classifier achieves higher accuracy in identifying and classifying different types of anomalies. The resulting identification model can be deployed in clinical settings to assist healthcare professionals in diagnosing and monitoring diseases more effectively. This approach improves diagnostic reliability and reduces the risk of misdiagnosis by leveraging advanced machine learning techniques tailored to medical imaging analysis.
5. The method according to claim 4 , wherein training the identification module comprises using clinical outcome.
This invention relates to a method for training an identification module in a medical or diagnostic system, specifically focusing on improving the accuracy of patient identification or diagnosis by incorporating clinical outcomes into the training process. The problem addressed is the need for more reliable and precise identification or diagnostic systems that can leverage real-world clinical data to enhance performance. The method involves training an identification module, which is part of a larger system designed to analyze medical data such as images, signals, or patient records. The identification module is trained using clinical outcomes, which are the actual results or conditions observed in patients after treatment or diagnosis. By incorporating these outcomes, the system learns to better distinguish between different conditions, reduce false positives or negatives, and improve overall diagnostic accuracy. The training process may involve machine learning techniques, where the identification module is adjusted based on feedback from clinical outcomes. This ensures that the system adapts to real-world scenarios, improving its ability to provide accurate and actionable insights. The method may also include preprocessing steps to prepare the clinical data for training, such as normalization or feature extraction, to enhance the learning process. By using clinical outcomes, the system avoids relying solely on theoretical or simulated data, leading to more practical and effective identification or diagnostic capabilities. This approach is particularly useful in fields like radiology, pathology, or personalized medicine, where accurate identification of patient conditions is critical for treatment decisions. The invention aims to bridge the gap between theoretical
6. The method according to claim 1 , further comprising providing a user interface for inputting a digital pathology image to the computing system and for outputting a display of detected pathological anomalies in the digital pathology image.
Digital pathology involves analyzing high-resolution images of tissue samples to detect and diagnose pathological anomalies. A key challenge is efficiently processing and visualizing these large, complex images to assist pathologists in identifying abnormalities. This invention addresses this problem by providing a computing system that processes digital pathology images to detect pathological anomalies and presents the results through an interactive user interface. The system includes a computing system configured to receive a digital pathology image, analyze the image to detect pathological anomalies, and generate a display of the detected anomalies. The analysis may involve image processing techniques such as segmentation, feature extraction, and machine learning-based classification to identify regions of interest. The system further includes a user interface that allows users to input digital pathology images and view the detected anomalies. The interface may highlight or annotate the anomalies in the image, providing pathologists with visual guidance to improve diagnostic accuracy and efficiency. The system may also support additional features, such as adjusting detection parameters or comparing multiple images, to enhance usability. This approach streamlines the workflow for pathologists by automating anomaly detection and providing an intuitive visualization tool.
7. The method according to claim 1 , wherein the pathological anomaly is an infection or an inflammation or a cancer tumor.
This invention relates to medical imaging and diagnostic systems, specifically for detecting and analyzing pathological anomalies in biological tissue. The technology addresses the challenge of accurately identifying and characterizing abnormalities such as infections, inflammations, or cancerous tumors within a patient's body using imaging techniques. The method involves capturing imaging data of a target region, processing the data to identify regions of interest, and applying analytical techniques to determine the presence and type of pathological anomaly. The system distinguishes between different types of anomalies, including infections, inflammations, and cancer tumors, by analyzing specific imaging markers or patterns associated with each condition. The invention improves diagnostic accuracy by reducing false positives and negatives, enabling earlier detection and more precise treatment planning. The method may integrate with existing imaging modalities such as MRI, CT, or ultrasound, enhancing their diagnostic capabilities. By automating the detection and classification process, the technology reduces the need for invasive procedures and speeds up the diagnostic workflow. The system can be used in clinical settings to assist healthcare professionals in making informed decisions about patient care.
8. The method according to claim 1 , wherein the machine learning module is a deep convolutional neural network.
A method for image processing using a deep convolutional neural network (CNN) to analyze and classify images. The method involves capturing an image using an imaging device, such as a camera, and preprocessing the image to enhance its quality and prepare it for analysis. The preprocessed image is then input into a deep convolutional neural network, which processes the image through multiple convolutional layers to extract features and identify patterns. The network applies convolutional filters to detect edges, textures, and other relevant features, followed by pooling layers to reduce dimensionality while retaining important information. The processed features are then passed through fully connected layers to generate a classification output, such as identifying objects, detecting anomalies, or recognizing patterns within the image. The method may also include post-processing steps to refine the results, such as filtering or thresholding, to improve accuracy. The deep convolutional neural network is trained on a dataset of labeled images to learn the relationships between image features and their corresponding classifications, enabling it to generalize and accurately classify new, unseen images. This approach is particularly useful in applications such as medical imaging, autonomous vehicles, and industrial inspection, where precise and automated image analysis is required.
9. The method according to claim 1 , wherein the machine learning algorithm is a deep convolutional neural network algorithm.
A system and method for image processing and analysis using machine learning, specifically a deep convolutional neural network (CNN), to improve accuracy and efficiency in tasks such as object detection, classification, or segmentation. The invention addresses the challenge of processing high-dimensional image data by leveraging the hierarchical feature extraction capabilities of CNNs, which automatically learn spatial hierarchies of features through convolutional layers. The method involves inputting an image into the CNN, where multiple convolutional layers apply filters to extract increasingly abstract features. These features are then processed through fully connected layers to produce a final output, such as a classification label or bounding box coordinates for detected objects. The CNN is trained on a labeled dataset to optimize its parameters, ensuring robust performance across diverse image inputs. This approach enhances computational efficiency and accuracy compared to traditional image processing techniques, making it suitable for applications in autonomous vehicles, medical imaging, and industrial automation. The use of deep learning eliminates the need for manual feature engineering, reducing development time and improving adaptability to new datasets.
10. The method according to claim 1 , wherein the at least one first image of the tissue slide and at least one second image of the tissue slide are images of a single tissue slide.
The invention relates to digital pathology and image analysis of tissue slides. The problem addressed is the need for accurate and efficient analysis of tissue samples, particularly when multiple images of the same slide are captured for different purposes, such as brightfield and fluorescence imaging. The invention provides a method for processing and analyzing images of a single tissue slide, where at least one first image and at least one second image of the same slide are obtained. These images may represent different imaging modalities, staining techniques, or magnification levels. The method involves aligning, registering, or comparing these images to extract meaningful biological or pathological information. This approach improves diagnostic accuracy by leveraging multiple perspectives of the same tissue sample, reducing errors from misalignment or inconsistencies between separate slides. The technique is particularly useful in applications requiring high-resolution analysis, such as cancer detection, tissue characterization, or biomarker quantification. By using a single slide, the method minimizes sample variability and ensures consistent data for analysis. The invention enhances workflow efficiency in pathology labs by reducing the need for multiple physical slides and streamlining image processing.
11. The method according to claim 1 , wherein said at least one fluorescent biomarker is a chemical compound that binds to a specific biological structure and can emit fluorescence when excited with light of a specific wavelength or an antibody that binds to a specific antigen and is labelled with a fluorescent dye.
This invention relates to a method for detecting biological structures using fluorescent biomarkers. The method addresses the challenge of accurately identifying and visualizing specific biological targets, such as proteins, cells, or other molecular structures, in biological samples. The method employs at least one fluorescent biomarker, which can be either a chemical compound or an antibody. The chemical compound binds to a specific biological structure and emits fluorescence when excited by light of a specific wavelength. Alternatively, the biomarker can be an antibody that binds to a specific antigen and is labeled with a fluorescent dye. Upon excitation, the fluorescent dye emits light, allowing the biological structure to be detected and analyzed. This approach enhances the specificity and sensitivity of biological detection, enabling precise identification of target structures in various applications, including medical diagnostics, biological research, and environmental monitoring. The method leverages the unique binding properties of chemical compounds or antibodies to ensure accurate targeting, while the fluorescent emission provides a clear and measurable signal for detection.
12. The method according to claim 1 , wherein said at least one fluorescent dye is an immunofluorescent dye.
This invention relates to a method for detecting or analyzing biological targets using fluorescent dyes, specifically immunofluorescent dyes. The method addresses the challenge of accurately identifying and quantifying biological molecules, such as proteins or cells, in a sample. Immunofluorescent dyes are used to label antibodies or other binding molecules that specifically target the biological molecules of interest. When excited by light, these dyes emit fluorescence, allowing the labeled targets to be visualized and measured. The method involves applying the immunofluorescent dye to a sample containing the biological targets. The dye binds to the targets via antibodies or other binding molecules, forming a fluorescent complex. The sample is then illuminated with light at a specific wavelength to excite the dye, causing it to emit fluorescence. The emitted fluorescence is detected and analyzed to determine the presence, quantity, or location of the biological targets in the sample. This approach enhances the specificity and sensitivity of detection, as immunofluorescent dyes provide strong, target-specific signals. The method is useful in applications such as medical diagnostics, biological research, and quality control in biopharmaceutical production. By using immunofluorescent dyes, the method improves the accuracy and reliability of detecting biological targets compared to non-specific or less sensitive techniques.
13. The method according to claim 1 , wherein said at least one fluorescent dyes is a fluorescent direct stain.
A method for analyzing biological samples involves using at least one fluorescent dye to label and detect target molecules within the sample. The fluorescent dye is a direct stain, meaning it binds directly to the target molecules without requiring additional labeling steps. This approach simplifies the staining process by eliminating the need for secondary reagents or conjugation steps, reducing time and potential variability in the assay. The method is particularly useful in applications such as microscopy, flow cytometry, or other analytical techniques where rapid and accurate detection of biological targets is required. The direct staining method enhances sensitivity and specificity by minimizing background noise and improving signal-to-noise ratios. This technique is applicable to various biological samples, including cells, tissues, or other biological materials, and can be used in research, clinical diagnostics, or industrial settings. The use of a fluorescent direct stain ensures efficient and reliable detection of target molecules, making it a valuable tool for biological analysis.
14. The method according to claim 1 , wherein said multiple fluorescent dyes are four different fluorescent dyes.
This invention relates to a method for analyzing biological samples using multiple fluorescent dyes to detect and quantify target molecules. The method addresses the challenge of accurately identifying and measuring multiple analytes in a single sample, which is critical in applications such as medical diagnostics, drug discovery, and environmental monitoring. Traditional methods often suffer from limitations in multiplexing capability, sensitivity, or specificity, leading to incomplete or unreliable results. The method involves labeling target molecules in a sample with four different fluorescent dyes, each emitting distinct wavelengths upon excitation. These dyes are selected to minimize spectral overlap, ensuring clear differentiation between signals. The labeled sample is then exposed to an excitation light source, and the emitted fluorescence is detected and analyzed. The system includes a detection device capable of resolving the distinct emission spectra of the four dyes, allowing simultaneous quantification of multiple analytes. The method may also incorporate signal processing techniques to enhance accuracy, such as background subtraction or spectral deconvolution. By using four distinct fluorescent dyes, the method enables high-throughput, multiplexed analysis with improved sensitivity and specificity compared to single-dye or lower-multiplex approaches. This allows for more comprehensive and efficient analysis of complex biological samples, reducing the need for multiple separate assays. The invention is particularly useful in applications requiring simultaneous detection of multiple biomarkers or genetic targets, such as cancer diagnostics, infectious disease testing, or genetic screening.
15. A method performed by a computing system for detecting pathological anomalies in a digital pathology image, comprising: providing a digital pathology image to the computing system; and analyzing the digital pathology image using an identification module arranged on the computing system, wherein the identification module uses a machine learning module to execute: recognizing an object containing an abnormal image pattern using an identification model loaded in said identification module; and identifying whether the abnormal image pattern corresponds to a pathological anomaly using the identification model, wherein providing the identification model comprises: providing a tissue slide with pathological anomalies; selecting at least one biomarker and at least one fluorescence dye, or at least one fluorescent biomarker, specific for a respective pathological anomaly; staining the tissue slide with the at least one biomarker and the at least one fluorescence dye or, the at least one fluorescent biomarker, wherein said tissue slide is stained with multiple fluorescent dyes which emit light at different wavelengths; generating at least one first digital image of the tissue slide; staining the tissue slide with a first histochemical dye; generating at least one second digital image of the same tissue slide, wherein the tissue slide is stained with a second histochemical dye, said second histochemical dye is H&E; annotating the at least one second image of the tissue slide by layering the at least one first digital image and the at least one second digital image over another; providing the annotated images to the machine learning module; and training the identification model using different morphological features of the pathological anomalies to obtain the classifier, and wherein the step of analyzing the digital pathology image comprises analyzing the multiple fluorescent dyes on the same tissue slide.
The invention relates to a method for detecting pathological anomalies in digital pathology images using machine learning. The method addresses the challenge of accurately identifying abnormal tissue patterns in pathology slides, which is critical for disease diagnosis and treatment planning. The system involves a computing system that processes digital pathology images by first staining a tissue slide with specific biomarkers and fluorescent dyes that target pathological anomalies. Multiple fluorescent dyes are used, each emitting light at different wavelengths, to capture distinct features of the anomalies. The stained slide is imaged to generate digital images. The slide is then stained with a first histochemical dye, and additional digital images are captured. A second histochemical stain, such as H&E (hematoxylin and eosin), is applied, and further imaging is performed. The digital images from the fluorescent and histochemical stains are layered and annotated to create a composite image. This annotated data is used to train a machine learning model, which learns to recognize abnormal patterns based on morphological features. The trained model is then deployed in an identification module to analyze new digital pathology images, detecting anomalies by examining the fluorescent and histochemical stains. The method improves diagnostic accuracy by leveraging multi-stain imaging and machine learning for comprehensive anomaly detection.
16. The method according to claim 15 , wherein providing a tissue slide comprises selecting a tissue slide from a patient where patient data and/or clinical treatment and/or clinical outcome are known.
This invention relates to a method for analyzing tissue samples, specifically focusing on selecting tissue slides from patients with known clinical data. The method involves obtaining a tissue slide from a patient where relevant patient data, clinical treatment details, and clinical outcomes are documented. This selection process ensures that the tissue samples used for analysis are associated with well-documented medical histories, allowing for more accurate correlations between tissue characteristics and clinical results. The method may include additional steps such as preparing the tissue slide for imaging, capturing digital images of the tissue, and analyzing these images to extract relevant biological or pathological information. The use of known clinical data enhances the reliability of the analysis, enabling better insights into disease progression, treatment effectiveness, and patient outcomes. This approach is particularly useful in medical research, diagnostics, and personalized medicine, where understanding the relationship between tissue characteristics and clinical data is critical. The method ensures that the tissue samples are representative and relevant, improving the accuracy and applicability of the findings.
17. The method according to claim 15 , wherein the at least one fluorescence dye is a fluorescent direct stain.
A method for analyzing biological samples involves using a fluorescent direct stain to detect and quantify specific components within the sample. The fluorescent direct stain binds directly to target molecules, such as nucleic acids or proteins, without requiring additional labeling steps. This approach simplifies the detection process by eliminating the need for secondary probes or amplification steps, reducing both time and cost. The method is particularly useful in applications like DNA quantification, protein analysis, or cell imaging, where rapid and accurate detection is critical. The fluorescent direct stain emits a detectable signal when excited by a specific wavelength of light, allowing for precise measurement of the target molecules. This technique enhances sensitivity and specificity, making it suitable for high-throughput screening and diagnostic assays. The method may also include steps for sample preparation, staining, and signal detection, ensuring reliable and reproducible results. By using a fluorescent direct stain, the method provides a streamlined and efficient way to analyze biological samples, improving workflow efficiency in research and clinical settings.
18. The method according to claim 15 , wherein the biomarker is a morphological biomarker and/or a cancer specific biomarker and/or another disease, different from cancer, specific biomarker.
This invention relates to a method for analyzing biological samples to detect and diagnose diseases, particularly cancer and other conditions, using biomarkers. The method involves identifying and quantifying specific biomarkers in a sample, which can include morphological biomarkers, cancer-specific biomarkers, or biomarkers associated with diseases other than cancer. Morphological biomarkers refer to structural or physical characteristics of cells or tissues that indicate disease presence or progression. Cancer-specific biomarkers are molecules or genetic markers uniquely or predominantly associated with cancer cells. The method may also detect biomarkers for non-cancerous diseases, allowing for broad diagnostic applications. The analysis can be performed using various techniques, such as imaging, molecular assays, or genetic sequencing, to provide accurate and early disease detection. The method aims to improve diagnostic precision by leveraging multiple biomarker types, enhancing the ability to distinguish between different diseases and their stages. This approach supports personalized medicine by tailoring treatments based on specific biomarker profiles. The invention addresses the need for more reliable and comprehensive diagnostic tools that can identify a wide range of diseases with high sensitivity and specificity.
19. The method according to claim 15 , wherein the histochemical dye are selected from a group comprising Haematoxylin, Eosin, van Gieson, Toluidine blue, Silver stain, Periodic acid-Schiff (PAS), Glycogen stain, Weigerts stain, Nissl stain, Golgi stain, Safranin, Oil Red, Prussian blue, Picro-Sirius Red, Mallary's trichome, Steiner Stain, Iron Hematoxylin and Fleugen stain.
This invention relates to histochemical staining techniques used in biological tissue analysis. The method involves selecting specific histochemical dyes to enhance the visualization of cellular and tissue structures under a microscope. The dyes are chosen from a group that includes Haematoxylin, Eosin, van Gieson, Toluidine blue, Silver stain, Periodic acid-Schiff (PAS), Glycogen stain, Weigerts stain, Nissl stain, Golgi stain, Safranin, Oil Red, Prussian blue, Picro-Sirius Red, Mallary's trichome, Steiner Stain, Iron Hematoxylin, and Fleugen stain. Each dye is selected based on its ability to bind to specific tissue components, such as proteins, carbohydrates, lipids, or nucleic acids, allowing for differential staining that highlights distinct structures. The method ensures precise and reproducible staining, aiding in diagnostic and research applications. The selection of dyes enables the identification of various tissue types, cellular structures, and pathological changes, improving the accuracy of histological analysis. The technique is particularly useful in medical diagnostics, where accurate tissue staining is critical for disease detection and treatment planning.
20. The method according to claim 15 , wherein annotating the at least one second image of the tissue slide comprises using the stained region of the at least one first image of the tissue slide as mask for cropping the at least one second image of the tissue slide.
This invention relates to digital pathology and image processing techniques for analyzing tissue slides. The problem addressed is the accurate annotation and segmentation of tissue regions in digital pathology images, particularly when multiple images of the same tissue slide are captured under different conditions (e.g., stained and unstained). The invention provides a method to improve the precision of tissue region identification by leveraging stained images as a reference for annotating unstained or differently processed images of the same tissue slide. The method involves capturing at least one first image of a tissue slide under a first imaging condition (e.g., stained with a specific dye) and at least one second image of the same tissue slide under a second imaging condition (e.g., unstained or differently stained). The stained regions in the first image are identified and used as a mask to crop or segment the corresponding regions in the second image. This ensures that only the relevant tissue areas are annotated in the second image, reducing noise and improving accuracy. The technique can be applied to various imaging modalities, including brightfield and fluorescence microscopy, and is useful for tasks such as automated tissue analysis, disease diagnosis, and research applications. The method enhances the reliability of digital pathology workflows by providing consistent and precise tissue region annotations across different imaging conditions.
21. The method according to claim 15 , wherein said at least one fluorescent biomarker is a chemical compound that binds to a specific biological structure and can emit fluorescence when excited with light of a specific wavelength or an antibody that binds to a specific antigen and is labelled with a fluorescent dye.
This invention relates to a method for detecting biological structures using fluorescent biomarkers. The method addresses the challenge of accurately identifying and visualizing specific biological targets in complex environments, such as biological samples or medical diagnostics, by leveraging fluorescent biomarkers that emit detectable light when excited by a specific wavelength. The method employs at least one fluorescent biomarker, which can be either a chemical compound that binds to a specific biological structure and emits fluorescence upon excitation or an antibody that specifically binds to a target antigen and is labeled with a fluorescent dye. The biomarker's fluorescence allows for precise detection and imaging of the biological structure when illuminated with the appropriate excitation light. This approach enhances sensitivity and specificity in biological assays, enabling applications in medical diagnostics, research, and imaging. The method may involve preparing a sample containing the biological structure, introducing the fluorescent biomarker, and exposing the sample to excitation light to induce fluorescence. The emitted fluorescence is then detected and analyzed to identify the presence and location of the biological structure. This technique is particularly useful in applications requiring high-resolution imaging or quantitative analysis of biological targets.
22. The method according to claim 15 , wherein said at least one fluorescent dye is an immunofluorescent dye.
This invention relates to a method for detecting or analyzing biological targets using fluorescent dyes, specifically immunofluorescent dyes. The method involves using at least one fluorescent dye to label biological targets, such as proteins, cells, or other biomolecules, for detection or analysis. The fluorescent dye emits light when excited by a specific wavelength, allowing the labeled targets to be visualized or quantified. The use of immunofluorescent dyes enhances specificity, as these dyes bind to targets via antibody-antigen interactions, reducing background noise and improving accuracy. The method may be applied in various fields, including medical diagnostics, biological research, and pharmaceutical development, where precise detection of biological targets is critical. The invention addresses the need for highly sensitive and specific detection techniques in complex biological samples, overcoming limitations of traditional staining methods. The immunofluorescent dye ensures strong signal-to-noise ratios, enabling reliable identification and analysis of targets in diverse applications.
23. The method according to claim 15 , wherein said multiple fluorescent dyes are four different fluorescent dyes.
This invention relates to a method for analyzing biological samples using multiple fluorescent dyes to detect and quantify specific components. The method addresses the challenge of accurately identifying and measuring multiple targets in a single sample, which is critical in applications such as medical diagnostics, environmental monitoring, and biological research. The method involves labeling different biological targets with distinct fluorescent dyes, each emitting light at a unique wavelength when excited by a light source. By using four different fluorescent dyes, the method enables simultaneous detection of four distinct targets within the same sample, improving efficiency and reducing the need for multiple separate assays. The dyes are selected to minimize spectral overlap, ensuring accurate differentiation between signals. The method may include steps such as preparing the sample, adding the fluorescently labeled probes, exciting the dyes with light, and detecting the emitted fluorescence using a suitable detection system. The use of four dyes allows for multiplexed analysis, providing a more comprehensive understanding of the sample composition in a single experiment. This approach enhances throughput and reduces costs compared to traditional single-target detection methods.
24. A method for annotating a tissue slide image, comprising: providing a tissue slide with pathological anomalies; staining the tissue slide with at least one biomarker and at least one fluorescence dye, or at least one fluorescent biomarker, specific for a respective pathological anomaly, wherein said tissue slide is stained with multiple fluorescent dyes which emit light at different wavelengths; generating at least one first digital image of the tissue slide; staining the tissue slide with a first histochemical dye; generating at least one second digital image of the same tissue slide, wherein the tissue slide is stained with a second histochemical dye, said second histochemical dye is H&E; annotating the at least one second image of the tissue slide by using a stained region of the at least one first image of the tissue slide; and providing the annotated images to the machine learning module and training the identification model using different morphological features of the pathological anomalies to obtain the classifier, wherein said step of annotating comprises layering the at least one first image of the tissue slide and the at least one second image of the tissue slide over another and using the stained region of the at least one first image of the tissue slide as mask for cropping the at least one second image of the tissue slide.
This invention relates to a method for annotating tissue slide images to improve pathological anomaly detection using machine learning. The method addresses the challenge of accurately identifying and classifying pathological features in tissue samples by combining fluorescence and histochemical staining techniques. The process begins with a tissue slide containing pathological anomalies, which is stained with at least one biomarker and at least one fluorescence dye, or a fluorescent biomarker, specific to the anomalies. The slide is stained with multiple fluorescent dyes that emit light at different wavelengths, allowing for distinct visualization of different pathological features. A first digital image of the stained slide is generated. The slide is then stained with a first histochemical dye, and a second digital image is captured after applying a second histochemical dye, specifically hematoxylin and eosin (H&E). The first and second images are aligned and overlaid, with the fluorescent-stained regions from the first image serving as a mask to annotate the corresponding areas in the second image. These annotated images are then provided to a machine learning module, where a classifier is trained using morphological features of the pathological anomalies. The method enhances annotation accuracy by leveraging the specificity of fluorescence staining to guide the interpretation of traditional H&E-stained images, improving the training of machine learning models for pathology detection.
25. The method according to claim 24 , wherein providing a tissue slide comprises selecting a tissue slide from a patient where patient data and/or clinical treatment and/or clinical outcome are known.
This invention relates to medical diagnostics, specifically methods for analyzing tissue samples to improve patient treatment and outcomes. The method involves selecting a tissue slide from a patient where detailed patient data, clinical treatment history, and clinical outcomes are known. This selection process ensures that the tissue sample is linked to comprehensive medical records, allowing for more accurate and context-aware analysis. The tissue slide is then processed to extract relevant biological information, which can be used to correlate specific tissue characteristics with treatment responses or disease progression. By leveraging known patient data, the method enables personalized insights that can guide future diagnostic or therapeutic decisions. The approach enhances the reliability of tissue-based diagnostics by ensuring that the samples are representative of well-documented cases, reducing variability and improving the predictive value of the analysis. This method is particularly useful in oncology, pathology, and precision medicine, where understanding the relationship between tissue features and clinical outcomes is critical for optimizing patient care.
26. The method according to claim 24 , wherein the at least one fluorescence dye is a fluorescent direct stain.
This invention relates to a method for analyzing biological samples using fluorescence microscopy, specifically addressing the challenge of accurately detecting and quantifying cells or cellular components in complex biological environments. The method involves staining the sample with at least one fluorescence dye, which is a fluorescent direct stain, to enhance contrast and visibility under fluorescence microscopy. The dye binds directly to cellular structures, such as nucleic acids or proteins, without requiring additional labeling steps, simplifying the staining process. The stained sample is then illuminated with excitation light at a specific wavelength, causing the dye to fluoresce. The emitted fluorescence is detected and analyzed to identify and quantify target cells or cellular components. The method may include additional steps such as sample preparation, imaging, and data processing to improve accuracy and reliability. The use of a fluorescent direct stain ensures efficient and specific staining, reducing background noise and improving detection sensitivity. This approach is particularly useful in applications such as cell counting, pathogen detection, and cellular imaging, where precise and rapid analysis is required. The method may be integrated into automated systems for high-throughput screening and analysis.
27. The method according to claim 24 , wherein the biomarker is a morphological biomarker and/or a cancer specific biomarker and/or another disease, different from cancer, specific biomarker.
This invention relates to a method for analyzing biomarkers to detect and diagnose diseases, particularly cancer and other conditions. The method involves identifying and evaluating specific biomarkers that indicate the presence or absence of disease. These biomarkers can be morphological, meaning they relate to the physical structure or appearance of cells or tissues, or they can be disease-specific, such as cancer-specific biomarkers that indicate the presence of cancerous cells. The method may also include biomarkers for diseases other than cancer, allowing for broader diagnostic applications. By analyzing these biomarkers, the method provides a way to detect and differentiate between various diseases, improving diagnostic accuracy and enabling earlier intervention. The approach leverages advanced analytical techniques to assess biomarker patterns, which can be used in clinical settings to guide treatment decisions. The method is designed to be adaptable, allowing for the integration of different types of biomarkers to enhance diagnostic precision.
28. The method according to claim 24 , wherein the second histochemical dye are selected from a group comprising Haematoxylin, Eosin, van Gieson, Toluidine blue, Silver stain, Periodic acid-Schiff (PAS), Glycogen stain, Weigerts stain, Nissl stain, Golgi stain, Safranin, Oil Red, Prussian blue, Picro-Sirius Red, Mallary's trichome, Steiner Stain, Iron Hematoxylin and Fleugen stain.
This invention relates to histochemical staining techniques used in biological tissue analysis. The problem addressed is the need for precise and reliable staining methods to enhance the visibility of specific tissue structures under a microscope, which is critical for diagnostic and research purposes. The invention describes a method for staining biological tissues using a combination of histochemical dyes, where a first dye is applied to the tissue to stain a first set of structures, and a second dye is applied to stain a second set of structures. The second dye is selected from a group of commonly used histochemical stains, including Haematoxylin, Eosin, van Gieson, Toluidine blue, Silver stain, Periodic acid-Schiff (PAS), Glycogen stain, Weigerts stain, Nissl stain, Golgi stain, Safranin, Oil Red, Prussian blue, Picro-Sirius Red, Mallary's trichome, Steiner Stain, Iron Hematoxylin, and Fleugen stain. The method ensures that the second dye does not interfere with the staining of the first set of structures, allowing for clear differentiation between different tissue components. This approach improves the accuracy of tissue analysis by providing distinct contrast between various cellular and extracellular elements, facilitating better diagnostic and research outcomes. The selection of the second dye depends on the specific structures to be highlighted, ensuring versatility in tissue staining applications.
29. The method according to claim 24 , wherein the pathological anomaly is an infection or an inflammation or a cancer tumor.
This invention relates to medical imaging and diagnostic techniques, specifically for detecting and analyzing pathological anomalies in biological tissue. The method involves using imaging data, such as from ultrasound, MRI, or CT scans, to identify and characterize abnormalities within the body. The technique focuses on distinguishing between different types of pathological conditions, including infections, inflammations, and cancerous tumors. By analyzing the imaging data, the method can determine the presence and nature of these anomalies, aiding in accurate diagnosis and treatment planning. The approach may involve processing the imaging data to extract features that differentiate between the various conditions, such as variations in tissue density, texture, or signal intensity. This allows for more precise identification of the type of anomaly present, improving diagnostic accuracy and enabling targeted medical interventions. The method can be applied in clinical settings to enhance the detection and characterization of pathological conditions, supporting better patient care and outcomes.
30. The method according to claim 24 , wherein the at least one first image of the tissue slide and the at least one second image of the tissue slide are images of a single tissue slide.
This invention relates to digital pathology and image analysis, specifically addressing the challenge of accurately analyzing and comparing multiple images of the same tissue slide to improve diagnostic accuracy. The method involves capturing at least one first image and at least one second image of a single tissue slide, where these images may be obtained under different conditions, such as varying magnification levels, staining techniques, or imaging modalities. The images are processed to extract relevant features, such as cellular structures, tissue patterns, or biomarkers, which are then compared to identify differences or correlations. This comparison helps in detecting abnormalities, verifying staining consistency, or validating diagnostic results. The method may also involve aligning the images to ensure accurate spatial correspondence between features in the different images. By analyzing multiple images of the same tissue slide, the technique enhances the reliability of pathological assessments, reduces errors from single-image analysis, and provides a more comprehensive understanding of tissue characteristics. The approach is particularly useful in clinical diagnostics, research, and quality control in pathology laboratories.
31. The method according to claim 24 , wherein said at least one fluorescent biomarker is a chemical compound that binds to a specific biological structure and can emit fluorescence when excited with light of a specific wavelength or an antibody that binds to a specific antigen and is labelled with a fluorescent dye.
This invention relates to a method for detecting biological structures using fluorescent biomarkers. The method addresses the challenge of accurately identifying and visualizing specific biological targets in a sample, which is critical for applications in medical diagnostics, biological research, and environmental monitoring. The method employs at least one fluorescent biomarker, which can be either a chemical compound that binds to a specific biological structure and emits fluorescence when excited by light of a specific wavelength, or an antibody that binds to a specific antigen and is labeled with a fluorescent dye. The biomarker is introduced into a sample containing the target biological structure or antigen. Upon excitation with light of the appropriate wavelength, the biomarker emits fluorescence, allowing the presence and location of the target to be detected and analyzed. This approach enhances sensitivity and specificity in biological detection, enabling precise identification of molecular interactions and structural features. The method may be used in various applications, including imaging, flow cytometry, and diagnostic assays, where accurate detection of biological targets is essential. The use of fluorescent biomarkers provides a versatile and highly sensitive tool for biological analysis, improving the reliability of results in research and clinical settings.
32. The method according to claim 24 , wherein said at least one fluorescent dye is an immunofluorescent dye.
This invention relates to a method for detecting or analyzing biological samples using fluorescent dyes, specifically immunofluorescent dyes. The method involves using at least one fluorescent dye to label biological targets, such as proteins, cells, or other biomolecules, for visualization or quantification. The immunofluorescent dye binds specifically to a target antigen or antibody, enabling precise detection through fluorescence microscopy or other imaging techniques. The method may include steps for preparing the sample, applying the fluorescent dye, and analyzing the resulting fluorescence signal. The use of immunofluorescent dyes enhances specificity and sensitivity, allowing for accurate identification and measurement of biological markers in research, diagnostics, or clinical applications. The technique is particularly useful in fields such as immunology, cell biology, and medical diagnostics, where precise detection of molecular interactions is critical. The method may be applied to various sample types, including tissue sections, cell cultures, or biological fluids, and can be adapted for automated or high-throughput analysis. The invention improves upon existing fluorescence-based detection methods by leveraging the specificity of immunofluorescent dyes, reducing background noise, and increasing detection accuracy.
33. The method according to claim 24 , wherein said multiple fluorescent dyes are four different fluorescent dyes.
This invention relates to a method for analyzing biological samples using multiple fluorescent dyes to detect and quantify specific components. The method addresses the challenge of accurately identifying and measuring multiple targets in a single sample, which is critical in applications such as medical diagnostics, environmental monitoring, and biological research. The method involves labeling different biological targets with distinct fluorescent dyes, each emitting light at a unique wavelength when excited by a light source. By using four different fluorescent dyes, the method enables simultaneous detection of four distinct targets within the same sample, improving efficiency and reducing the need for multiple separate assays. The dyes are selected to minimize spectral overlap, ensuring accurate differentiation between signals. The method may include steps such as preparing the sample, adding the fluorescently labeled probes or antibodies, exciting the dyes with a light source, and detecting the emitted fluorescence using a suitable detection system, such as a fluorescence microscope or flow cytometer. The use of four dyes allows for multiplexed analysis, providing comprehensive data from a single experiment. This approach enhances throughput, reduces sample consumption, and improves the reliability of results by minimizing variability between separate assays. The method is particularly useful in high-throughput screening and complex biological studies where multiple analytes must be detected simultaneously.
34. A system for detecting pathological anomalies in a digital pathology image, comprising: a digital pathology image receiving module; and a digital pathology image analyzing module; wherein the digital pathology image analyzing module comprises an identification module, wherein the identification module comprises a machine learning module and an identification model, wherein the machine learning module comprises: a recognizing module for recognizing an object containing an abnormal image pattern using the identification model; and an identifying module for identifying whether the abnormal image pattern corresponds to a pathological anomaly using the identification model, wherein the identification model is provided by using a machine learning algorithm trained on a plurality of annotated digital pathology images; generating at least one first digital image of a tissue slide comprising stained morphological features of the pathological anomaly, wherein the tissue slide is stained with: a first histochemical dye, or at least one biomarker and at least one fluorescence dye, or at least one fluorescent biomarker; generating at least one second digital image of the same tissue slide, wherein the tissue slide is stained with a second histochemical dye, said second histochemical dye is H&E; annotating the at least one second digital image of the tissue slide by layering the at least one first digital image and the at least one second digital image over another; and providing the annotated images to the machine learning module and training the identification model using different morphological features of the pathological anomalies to obtain the classifier, wherein said tissue slide is stained with multiple fluorescent dyes which emit light at different wavelengths; and wherein the digital pathology image analyzing module is configured to analyze the multiple fluorescent dyes on the same tissue slide.
The system detects pathological anomalies in digital pathology images by analyzing stained tissue slides. The system addresses challenges in accurately identifying abnormalities in tissue samples, which is critical for diagnostic accuracy in medical pathology. The system includes a module to receive digital pathology images and an analyzing module that uses machine learning to detect and classify anomalies. The analyzing module contains an identification module with a machine learning component and an identification model. The machine learning module recognizes objects with abnormal patterns and identifies whether these patterns correspond to pathological anomalies using a pre-trained model. The model is trained on annotated digital pathology images, where the annotations are generated by overlaying images of the same tissue slide stained with different dyes. The system uses multiple staining techniques, including histochemical dyes, biomarkers with fluorescence dyes, or fluorescent biomarkers, to capture different morphological features. The tissue slides are stained with multiple fluorescent dyes emitting light at different wavelengths, allowing the system to analyze these dyes on the same slide. This multi-stain approach enhances the detection of pathological anomalies by providing comprehensive morphological data for training the machine learning model. The system improves diagnostic accuracy by leveraging advanced imaging and machine learning techniques to analyze complex tissue samples.
35. The system according to claim 34 , wherein the identification model comprises a classifier to classify the abnormal image pattern in accordance with at least two classes; and a determining module for determining whether the abnormal image pattern corresponds to a pathological anomaly based on the classification.
This invention relates to medical imaging systems designed to detect and classify abnormal patterns in medical images, such as those from X-rays, MRIs, or CT scans. The system addresses the challenge of accurately identifying and categorizing pathological anomalies in medical imaging data, which is critical for early diagnosis and treatment planning. The system includes an identification model that processes medical images to detect abnormal patterns. The model incorporates a classifier to categorize these patterns into at least two distinct classes, enabling finer-grained analysis. Additionally, a determining module evaluates the classified patterns to assess whether they indicate a pathological anomaly, providing a more precise diagnostic output. The system enhances diagnostic accuracy by reducing false positives and improving the specificity of anomaly detection, which is particularly valuable in high-stakes medical applications where misdiagnosis can have severe consequences. The invention leverages machine learning techniques to improve the reliability of automated medical image analysis, supporting healthcare professionals in making informed decisions.
36. The system according to claim 34 , further comprising an identification model providing module, wherein the identification model providing module comprises a training module for training a machine learning algorithm based at a plurality of annotated digital pathology images.
This system relates to digital pathology and the use of machine learning for analyzing pathology images. The problem addressed is the need for accurate and automated identification of features in digital pathology images, which is critical for diagnostic and research applications. The system includes a module that provides an identification model trained on annotated digital pathology images. This module contains a training component that trains a machine learning algorithm using a dataset of labeled pathology images. The training process involves processing the annotated images to extract relevant features and patterns, enabling the algorithm to recognize and classify structures or conditions in new, unlabeled pathology images. The trained model can then be deployed to assist pathologists by automating the detection of abnormalities, reducing human error, and improving diagnostic efficiency. The system may also include other components, such as image preprocessing modules, feature extraction tools, and user interfaces for model deployment and result visualization. The overall goal is to enhance the accuracy and speed of pathology analysis through machine learning-driven automation.
37. The system according to claim 34 , further comprising: a user interface comprising: an input module for providing a digital pathology image to the system; and an output module for displaying detected pathological anomalies in the digital pathology image.
This invention relates to digital pathology systems designed to analyze and detect pathological anomalies in medical images. The system processes digital pathology images, such as whole-slide images, to identify and highlight abnormalities like tumors, lesions, or other irregularities. The system includes a user interface with an input module for uploading or selecting digital pathology images and an output module for visualizing the detected anomalies. The output module may display annotations, markers, or heatmaps to indicate the location and severity of anomalies within the image. The system may also include image preprocessing, segmentation, and classification algorithms to enhance accuracy. The user interface allows pathologists or medical professionals to interact with the system, facilitating faster and more accurate diagnosis. The invention aims to improve diagnostic efficiency by automating anomaly detection while providing clear visual feedback to users. The system may integrate with existing pathology workflows, supporting both standalone and collaborative analysis. The user interface ensures seamless integration of digital pathology tools into clinical practice, reducing manual review time and enhancing diagnostic confidence.
Unknown
February 25, 2020
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.